J. Phys. Chem. C 2008, 112, 8481–8485
8481
Hydrogen Sorption Behavior of the ScH2-LiBH4 System: Experimental Assesment of Chemical Destabilization Effects Justin Purewal,*,† Son-Jong Hwang,‡ Robert C. Bowman, Jr.,§ Ewa Ro¨nnebro,⊥ Brent Fultz,† and Channing Ahn† DiVisions of Engineering and Applied Science, and Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125, Energy Systems Department, Sandia National Laboratories, LiVermore, California 94551, and Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109 ReceiVed: January 17, 2008; ReVised Manuscript ReceiVed: February 27, 2008
The hydrogen storage reaction ScH2 + 2LiBH4 f ScB2 + 2LiH + 4H2 (8.91 wt %), based on the thermodynamic destabilization of LiBH4, is predicted to have a reaction enthalpy of ∆H300K ) 34.1 kJ/mol H2. The isothermal kinetic desorption behavior in this system was measured. At temperatures up to 450 °C, less than 5 wt % H2 is released, which is only half of the theoretical capacity. Powder X-ray diffraction data indicate that LiBH4 has decomposed into LiH in the final desorption product, but the data provide no evidence that ScH2 has participated in the reaction. Magic angle spinning NMR (MAS NMR) results do not show that the expected ScB2 equilibrium product phase has formed during desorption. While the addition of 2 mol % TiCl3 catalyst does improve desorption kinetics at 280 °C, it does not otherwise assist the destabilization reaction. The calculated reaction enthalpy suggests that this system should be of interest at moderate temperatures, but the large heats of formation of the reactant phases in this system appear to play a critical role in determining overall kinetics. Furthermore, the formation of a Li2B12H12 intermediate phase was determined by MAS NMR, which is an undesirable stable product if reaction reversibility is to be accomplished. Introduction Technologically relevant solid-state hydrogen storage materials must have gravimetric storage capacities significantly exceeding 6 wt % to meet the requirements promoted by the U.S. Department of Energy for applications such as fuel-cell powered vehicles.1 Complex metal hydrides of the light elements have been the subject of intense research as materials meeting these strict gravimetric and volumetric criteria. One promising material is LiBH4, which can release a maximum of 13.8 wt % hydrogen upon thermal decomposition:
LiBH4 f LiH + B + 3/2H2
(1)
However, LiBH4 is too stable to release and reabsorb hydrogen at ambient temperature and pressure. The standard reaction enthalpy for the decomposition of LiBH4 in the liquid state is 61.1 kJ/mol H2, which translates to a 1 bar equilibrium pressure at 400 °C.2,3 Chemical destabilization of LiBH4 offers a versatile way to lower the desorption enthalpy.3–8 Recent work has focused on adding an additional metal (M) or metal hydride (MH2) species which reacts to form a stable metal boride phase, lowering the overall reaction enthalpy.7,8 Vajo et al. demonstrated that adding MgH2 to LiBH4 lowered the dehydrogenation enthalpy by 25 kJ/mol and also made the system reversible.3 Motivated by such successes, a comprehensive evaluation of potential chemically * To whom correspondence should be addressed. E-mail: purewal@ caltech.edu. † Division of Engineering and Applied Science, California Institute of Technology. ‡ Division of Chemistry and Chemical Engineering, California Institute of Technology. § Jet Propulsion Laboratory. ⊥ Sandia National Laboratories.
destabilized metal hydride systems was initiated, using density functional theory (DFT) to evaluate formation and reaction enthalpies.10,11 A significant result from this study was the identification of ScH2 as a destabilizing agent for LiBH4 with ideal thermodynamics.9 The identified decomposition pathway,
ScH2 + 2LiBH4 f ScB2 + 2LiH + 4H2
(2)
has a calculated reaction enthalpy at 300 K of ∆H300K ) 34.1 kJ/mol after the zero point energy correction.9 This system can release 8.91 wt % H2 upon completion, and calculations predict the system to have an equilibrium pressure of 1 bar at around 57 °C It is important to note, however, that first principles DFT calculations provide no information about either the reaction mechanism or the activation barriers. It is likely that the rate limiting steps would involve the breakdown of ScH2 (∆Hf ) 200 kJ/mol)10 and the formation of ScB2 (∆Hf ) 248 kJ/mol),10 both of which are very stable compounds. While DFT has proved accurate in reproducing experimental H2 equilibrium pressures for the destabilized MgH2-LiBH4 system,10 it is uncertain whether it will work as well for a reaction system containing reactant and product phases that are considerably more stable. High energy ball-milling can improve reaction kinetics to some degree by reducing diffusion distances and increasing the concentration of reaction interfaces, but it does not necessarily mitigate the large activation barriers to ScH2 and ScB2 decomposition. The aim of the present work is to experimentally investigate the destabilized LiBH4-ScH2 system and evaluate the importance of kinetic barriers in determining hydrogen absorption and desorption behavior. Hydrogen sorption properties are studied with isothermal kinetics measurements, while reaction products are characterized by powder
10.1021/jp800486n CCC: $40.75 2008 American Chemical Society Published on Web 05/07/2008
8482 J. Phys. Chem. C, Vol. 112, No. 22, 2008
Purewal et al.
X-ray diffraction (XRD), magic angle spinning nuclear magnetic resonance (MAS NMR), and Raman spectroscopy. Experimental Details ScH2 was synthesized by heating small scandium ingots (99.9%, Stanford Materials Corporation) in ultrahigh purity (UHP) grade H2 at 350 °C. The resulting formation of scandium hydride was rapid and highly exothermic. Hydrogenation was allowed to proceed for 3 h to ensure equilibration. Volumetric analysis and X-ray diffraction both indicate that the reaction product is ScH2. Roughly 10 g of a 2:1 molar mixture of LiBH4 (95% purity, Sigma Aldrich) and ScH2 was loaded under argon atmosphere into a sealed 80 mL stainless steel vessel containing five 0.5 in. diameter stainless steel balls. The vessel was sealed with a rubber gasket in a high purity argon glovebox. The argonfilled vessel was then loaded into a Fritsch-pulverizette 6 planetary mill. Different mixtures were milled at 400 rpm for either 1 or 10 h. For possible catalytic enhancement of the reaction rates,3,7 2 mol % of TiCl3 (Sigma Aldrich) was added to one sample prior to milling. All samples were handled in argon atmosphere gloveboxes (Vacuum Atmospheres Corporation) to prevent air and moisture contamination. Isothermal kinetic desorption measurements were conducted on a custom built stainless steel Sieverts apparatus equipped with a capacitance pressure transducer (MKS Instruments). In a typical experiment, a 1 g sample was loaded under argon atmosphere into a 14 mL stainless steel reactor assembly sealed with a 0.5 in. Swagelok VCR filter gasket connection. The reactor was then mounted on the Sieverts system and enclosed in a thick, tight-fitting aluminum collar to maintain temperature uniformity. The furnace setup included a Watlow band heater fastened around the aluminum collar, two thermocouples inserted into holes in the aluminum collar, and an Omega temperature controller. The quantity of desorbed hydrogen was determined by volumetric measurements. Weight percents are reported with respect to the fully hydrogenated samples, excluding the weight of the catalyst. Reabsorption was not attempted due to difficulties in extracting the desorption products. Powder X-ray diffraction (XRD) patterns were acquired on a PANalytical X’pert PRO X’celerator diffractometer using Cu KR radiation (λ ) 1.5418 Å). All samples were sealed in 1.0 mm glass capillary tubes in an argon glovebox. Raman spectroscopy measurements were performed on a Renishaw M1000 Micro-Raman spectrometer, operating at 1 cm-1 spectral resolution with a 514.5 nm argon laser, 1800 nm grating, and 20× LWD objective. Solid-state magic angle spinning nuclear magnetic resonance (MAS NMR) measurements were performed using a Bruker Avance 500 MHz spectrometer equipped with a Bruker 4 mm boron-free cross-polarization (CP) MAS probe. The spectral frequencies were 500.23, 160.5, 121.6, and 73.6 MHz for the 1H, 11B, 45Sc, and 6Li nuclei, respectively. Samples were loaded into 4 mm ZrO2 rotors, and each was sealed with a tight fitting kel-F cap in an argon glovebox and spun at 13 kHz. The onedimensional (1D) 11B, 45Sc, and 6Li MAS NMR spectra were acquired after a short (∼0.5 µs) single pulse (i.e.,
